Complexities of ascorbate as a reducing agent

Prévia do material em texto

Inorg. Chem. 1981, 20, 4449-4452 4449
with the frequency of the exchange of the metal ion between
sites A and A'.
In order to make the concept as clear as possible, we want
to extend the designation and we define the following sub-
classes: Isoalterdentate ligands can bind a metal ion at two
or more symmetrically equivalent positions. If one of the sites
is occupied, the others are still available for coordination. The
number of sites can be included as a prefix. Nin" (I) and all"
(II) are accordingly diisoalterdentate ligands. Sometimes, it
depends on the nature of the metal ion whether a ligand acts
as a chelate or as an alterdentate. In most cases en is a chelate,
whereas for Ag+ it is an alterdentate in the complex Ag(en)2+,
owing to the linear coordination of the silver ion.1 234 S-Triazine
(III) is triisoalterdentate, and the cyclopentadienyl anion (IV)
is pentaisoalterdentate in some (n1-C5H5)M species.
(Ill) (IV)
There is another type of ligand, exemplified by (V),
(Va) (Vb) (Vc)
which can offer equivalent coordination sites. The metal can
change from A to A' to A" (Va to Vc) only if the ligand itself
rearranges simultaneously, however, and the same type of
coordination sites is no longer available if one metal is bound.
This class is called anisoalterdentate, and TPTZ is therefore
trianisoalterdentate. AG° for the rearrangement is still 0.
Ambidentate ligands offer nonequivalent coordination sites
to the metal, and, generally, AG° ^ 0 for the exchange be-
tween sites. Of the three different isotopic species 1SN14N14N",
14N15N14N", and 14N14N14N", having natural abundances of
0.74%, 0.37%, and 98.89%, the first is ambidentate and the
second and the third are isoalterdentate.
The two types can be combined and the phenazine (VI)
(VI)
derivative with five sites AA'BB'C, of which AA' and BB' are
pairwise equivalent, can be disignated as isoalterambidentate.
The proposed classification scheme emphasizes a possible
nonrigidity in metal complexes, which has not been very often
considered, so far, in coordination chemistry. Many alter-
dentate ligands are known, and many more could be designed.
Experimental observation of rearrangements will, in most
cases, be possible by NMR spectroscopy with diamagnetic
ligands or with ESR in the case of radical ligands if the metals
are sufficiently labile. For substitution-inert complexes, other
methods such as isotopic labelling might be used.
Complicated polyfunctional ligands in biochemical systems
could, in principal, exchange metal ions between several
equivalent coordination sites, and alterdenticity may play an
important role in some bioinorganic reactions, including the
(4) Magyar, B.; Schwarzenbach, G. Acta Chem. Scand., Ser. A 1978, A32,
943.
transport of metal ions over considerable distances.
Acknowledgment. This work is supported by the Swiss
National Science Foundation. The author thanks reviewers
for useful comments.
Institute of Inorganic Chemistry A. von Zelewsky
University of Fribourg
CH-1700 Fribourg, Switzerland
Received February 20, 1981
The Complexities of Ascorbate as a Reducing Agent
Sir:
Aqueous solutions of ascorbic acid (H2A, vitamin C) have
been widely used by inorganic chemists because of the con-
venience of ascorbate as a reducing agent.1"3 We have re-
cently utilized this reductant as a source of reducing equiva-
lents in a multicomponent system that promotes the photo-
reduction of water.4 In connection with this work information
on the redox characteristics of the various species generated
in the oxidation of ascorbate in water was sought. Here a
summary of these findings is presented.
The structures I, II, and V shown in Figure 1 are generally
given5 for ascorbic acid, ascorbate ion (HA"), and dehydro-
ascorbic acid (A), respectively, but it has recently been es-
tablished that VI is actually that appropriate to dehydro-
ascorbic acid in aqueous solutions.6 On the basis of EPR
results Schuler and co-workers7 have concluded that IV is the
structure of ascorbate radical (A"·), the one-electron oxidation
product of ascorbic acid or of ascorbate ion; its structure is
invariant in the pH range 0-13, but at lower pH, A"· pro-
tonates to give III.7
Ascorbic acid is a weak acid,8 while the radical HA· is a
very strong acid.7 Observations by Ball suggest that dehy-
droascorbic acid undergoes proton loss with a pof ~8.9
Data relevant to the various protonation equilibria are sum-
marized in Table I.
Thermodynamic data bearing on redox equilibria between
the various species come from several sources. By recourse
to mediators Ball performed potentiometric measurements on
the ascorbic acid/dehydroascorbic acid couple (eq 1) in
A + 2H+ + 2e" = H2A (1)
aqueous solutions ranging from pH 1 to 8.57 at 30 °C.9 The
reduction potential for the couple was determined to be +0.39
V in 1 N acid at 30 °C.9 The A/H2A couple was found to
be chemically reversible in a practical sense up to about pH
7; above pH 7, the decomposition of the oxidized form became
rapid, with the half-life for its decomposition decreasing from
~40 min at pH 6.7 to ~0.5 min at pH 8.6. Foerster, Weis,
and Staudinger10 used an ESR method to evaluate the con-
(1) Kustin, K.; Toppen, D. L. Inorg. Chem. 1973,12,1404. Rickman, R.
A.; Sorenson, R. L.; Watkins, K. O.; Davies, G. Ibid. 1977,16, 1570.
(2) Harris, F. L.; Toppen, D. L. Inorg. Chem. 1978,17. 74. Oxley, J. C.;
Toppen, D. L. Ibid. 1978, 17, 3119.
(3) Jwo, J. J.; Haim, A. J. Am. Chem Soc. 1976, 98, 1172.
(4) Brown, G. M.; Brunschwig, B. S.; Creutz, C.; Endicott, J. F.; Sutin, N.
J. Am. Chem. Soc. 1979, 101, 1298.
(5) Lehninger, A. L. “Biochemistry. The Molecular Basis of Cell Structure
and Function”; Worth: New York, 1970; p 25.
(6) Pfielstucker, K.; Marx, F.; Bockisch, M. Carbohydr. Res. 1975,45,269.
(7) Laroff, G. P.; Fessenden, R. W.; Schuler, R. H. J. Am. Chem. Soc.
1972, 94, 9062.
(8) Martell, A. E.; Smith, R. M. “Critical Stability Constants”; Plenum
Press: New York, 1979; Vol. 3, p 264.
(9) Ball, E. G. J. Biol. Chem. 1937, 118, 219.
(10) Foerster, G. v.; Weis, W.; Staudinger, H. Liebigs Ann. Chem. 196$, 690,
166.
0020-1669/81 /1320-4449S01.25/0 © 1981 American Chemical Society
D
ow
nl
oa
de
d 
vi
a 
U
N
IV
 O
F 
SA
O
 P
A
U
LO
 o
n 
M
ar
ch
 1
8,
 2
01
9 
at
 1
9:
02
:2
4 
(U
TC
). 
Se
e 
ht
tp
s:/
/p
ub
s.a
cs
.o
rg
/sh
ar
in
gg
ui
de
lin
es
 fo
r o
pt
io
ns
 o
n 
ho
w
 to
 le
gi
tim
at
el
y 
sh
ar
e 
pu
bl
ish
ed
 a
rti
cl
es
. 
4450 Inorganic Chemistry, Vol. 20, No. 12, 1981 Correspondence
ha·, m
:ohHCOH
I
CH,0H
 , 
ch2oh
a· , m
Figure 1. Structure of ascorbic acid (I, H2A), ascorbate ion, (II, HA"),
protonated ascorbate radical (III, HA·), ascorbate radical (IV, A"·),
and dehydroascorbic acid (V and VI, A).
Table I. Dissociation Constants for Ascorbic Acid, Ascorbate
Radical, and Dehydroascorbic Acid
equilibrium pK& ref
Table II. Reduction Potentials for Ascorbate Species in
Aqueous Solution0
couple £°, V
(i) A + 2H+ + 2e" = H2A +0.40
(ii) A + H+ + 2e~ = HA" + 0.28
(iii) A + 2e" = A2" -0.05
(iv) A + e" = A"· -0.16, -0.12°·°
(v) A"· + e" = A2" +0.05 ,a’b +0.015°
(vi) HA· + e" = HA" +0.70,°·6 +0.66°·°
0 Dissociation constants summarized in Table I have been used
in these calculations. The value chosen as E° for the A/H2A
couple (entry i) represents a compromise between 0.39 V deter-
mined by Ball9 at 30 °C and other values of ~0.41 V at 25 °C.13
6 The one-electron reduction potentials were calculated from the
ascorbate radical disproportionation equilibrium data obtained by
Foerster et al.‘° and evaluated by Bielski." ° Based on the
 "·/ 2~ E° value determined by Steenken and Neta at pH 13.5
from thereaction of ascorbate with catechol.’2
idation reactions. Pelizetti and co-workers have compiled
extensive data on outer-sphere oxidation of HA" and H2A by
such oxidants as FeL33+ (L = a 2,2'-bipyridine or 1,10-
phenanthroline derivative) and found the rate law for these
reactions in —-0.1—1 M acid to be of the form given by eq 3.14·15
d[FeL32+]
~-d7
- = fc/[FeL33+] [HA~] (3)
(For some very strong oxidants, a term first order in oxidant
and first order in H2A was also important but will not be
discussed further here.) The rate law eq 3 was interpreted
in terms of Scheme I where the first step is rate determining
Scheme I
Ascorbic Acid
H2A^H+ + HA" p>f1 = 4.0‘2 8
HA" ^H* + A2" pK2 = 11.3°
Ascorbate Radical
HA· =* H+ + A"· pK = -0.45 7
Dehydroascorbic Acidb
A ^ H+ + (A - H)" pK = ~8 9
0 At 25 °C and 0.1 M ionic strength. 6 The formula (A - H)~
denotes the conjugate base of A.
centration of ascorbate radical resulting when ascorbate and
dehydroascorbic acid are mixed at 25 °C, pH < 6.4 (ionic
strength » (3 X 10"3)-(3 X 10"') M). Bielski has recently
used their data to evaluate an equilibrium constant of 2 X 10"15
M for eq 2.11 The results obtained by Ball and by Bielski may
A + HA" ^ 2A"· + H+ (2)
be combined with values determined for the acidity constants
of the various species (Table I) to give reduction potentials
for several one- and two-electron couples important to the
reactions of ascorbate species. Data for these couples are
presented in Table II. Also presented in Table II are estimates
based on the E° value obtained by Steenken and Neta for the
A"·/A2" couple.12 As is evident from the table, values obtained
for the radical couples on the basis of EPR and pulse radiolysis
methods are in reasonably good agreement. The 40-mV
difference between the two data sets may reflect the differing
media used in the two experiments and/or the accumulated
effects of experimental errors.
Having addressed the thermodynamic aspects of the as-
corbate system, we turn now to the kinetics of ascorbate ox-
HA" + FeL33+ ^ HA· + FeL32+ kh slow
HA· + FeL33+ — A + H+ + FeL32+ k2
so that k' = 2k¡. Values determined for ki ranged from 107
to >109 M"1 s"1. Added FeL32+ did not alter the observed
kinetics, which were well-behaved under pseudo-first-order
conditions ([H2A] > 10[FeL33+]). The possible effects of HA·
deprotonation (eq 4) and ascorbate radical disproportionation
(eq 5) were not addressed. Either or both could however be
HA· ^ H+ + A"· (4)
A". + A"· ——*· HA" + A (5)
relevant since A"· is half protonated only at ~2.8 M H+ (the
acidity range encompassed was 0.1-1.0 M), and the effective
second-order rate constant for eq 5 is probably >108 M"1 s"1
in the acidity range used.11 In a very recent study of the
oxidation of ascorbate in acidic solution both forward and
reverse rate constants for ascorbate radical production were
directly measured16 with the oxidant used being the chloro-
promazine cation radical (C1PMZ+).
Scheme II
HA" + Fein(cyt c) -*· HA· + Fen(cyt c) kt
HA· ^ H+ + A"·
A"· + Fem(cyt c) — A + Feu(cyt c) k2
A2" + Fem(cyt c) A"· + FeH(cyt c) k2
2A". + H+ -* HA" + A fc4
By contrast, studies of the oxidation of ascorbate by the
milder oxidant ferricytochrome c (Fem(cyt c), E° = 0.26 V17)
(11) Bielski, B. H. J. Adv. Chem. Ser., in press.
(12) Steenken, S.; Neta, P. J. Phys. Chem. 1979, 83, 1134.
(13) Clark, W. M. “Oxidation Potentials of Organic Systems"; Williams and
Wilkins: Baltimore, Md., 1960.
(14) Pelizetti, E.; Mentasti, E.; Pramauro, E. Inorg. Chem. 1976,15, 2898.
(15) Pelizetti, E.; Mentasti, E.; Pramauro, E. Inorg. Chem. 1978, 17, 1181.
(16) Pelizetti, E.; Meisel, D.; Mulac, W. A.; Neta, P. J. Am. Chem. Soc.
1979, 101, 6954.
Correspondence Inorganic Chemistry, Vol. 20, No. 12, 1981 4451
in neutral solution have implicated the mechanism given in
Scheme II,18 where kx ~ 1 M"1 s_1,18b k2 ~ 104 M"1 s-1,18 and
k2 ~ 5 X 105 M"1 s-1.18 Depending upon the conditions used,
the ascorbate radical decayed through oxidation by Fem(cyt
c) or through the disproportionation reaction eq 5 for which
the effective value of k4 was (1-2) X 105 M"1 s"1. Both Scheme
I and Scheme II lead to the stoichiometry given in eq 6.
HA' + 2Feni — A + H+ + 2Fen (6)
The results of these studies illustrate some of the com-
plexities of ascorbate as a reductant. Depending upon the pH
and the properties of the oxidant, H2A, HA", and/or A2" may
be the kinetically significant reducing agent(s). The cyto-
chrome c work shows that, depending upon the H+ ion and
reactant concentrations, the ascorbate radical may be con-
sumed via eq 5 or through reaction with the added oxidant.
On the other hand, the assumptions used in treating the FeL33+
kinetics require that oxidation of HA· or A · by FeL33+ pre-
dominate. While these assumptions are probably valid under
many conditions, a recent reinvestigation16 of the chloro-
promazine cation reduction by ascorbate19 suggests that ox-
idation of ascorbate radical can be relatively slow in acid as
well as in neutral solution. The reinvestigation indicates that,
in 0.25-1.0 M H+, an equilibrium between C1PMZ+, ascorbic
acid, C1PMZ, and the ascorbate radical is rapidly set up and
that reaction of the ascorbate radical with C1PMZ+, itself, or
impurities is rate determining.16 Clearly the kinetics of as-
corbate reactions in acid (as well as in neutral solution) merit
further attention. In any case the above studies do provide
some information bearing on the general kinetic reactivity of
the various ascorbate couples. This subject is considered in
the following paragraphs.
It is worthwhile to discuss the reactivity of species toward
outer-sphere redox reactions in terms of the driving force for
the electron-transfer reaction and the intrinsic electron-transfer
barrier associated with each of the reactants. This dichotomy
is embodied in the Marcus cross-relation,20 kn = (knk22Kny/2,
where ku and Ki2 are respectively the rate and equilibrium
constant for the electron-transfer reaction and kn and k22 are
self-exchange rate constants for the reactant species involved
(see below). In Table II are one-electron E° values, which
may be used to evaluate Kl2 for outer-sphere reactions in-
volving ascorbate species. The next focus of discussion is the
intrinsic electron-transfer barrier (reflected in the magnitude
of fcu) relevant to each of the ascorbate couples. First the
ascorbate/ascorbate radical self-exchange processes eq 7-9
HA' + HA· ^ HA· + HA" (7)
H2A + 2 +· — 2 +· + H2A (8)
A2" + A · ^ A · + A2’ (9)
are considered. Pelizetti and co-workers have attempted to
systematically characterize the reactivity of the HA /HA· and
H2A/H2A+· couples with outer-sphere oxidants.14,15 From the
free energy dependences of HA" and H2A oxidation rate
constants with characterized outer-sphere reagents E° esti-
mates for these couples were extracted by assuming that the
intrinsic barriers to electron transfer are not large for these
couples. By this means an E° estimate of 0.85-1.0 V was
obtained for the HA-/HA" couple. Since that estimate is at
(17) Margalit, R.; Schejter, A. Eur. J. Biochem. 1973, 32, 492.
(18) (a) Yamazaki, I. J. Biol. Chem. 1962, 237, 224. (b) Bielski, B. H. J.;
Richter, H. W.; Chan, P. C. Ann. N.Y. Acad. Sci. 1975, 258, 231, and
unpublished work.
(19) Klein, N. A.; Toppen, D. L. J. Am. Chem. Soc. 1978, 100, 4541.
(20) Marcus, R. A. Amu. Rev. Phys. Chem. 1966, 15, 155. See the fol-
lowing for comments on the application and reliability of this analysis:
Bottcher, W.; Brown, G. M.; Sutin, N. lnorg. Chem. 1979, 18, 1447.
Chou, M.; Creutz, C.; Sutin, N. J. Am. Chem. Soc. 1977, 99, 5615.
least 150 mV more positive than the value derived from the
thermodynamic experiments (Table II), it is likely that the
driving force for the oxidation reactions was underestimated.
Consequently, it islikely that the intrinsic electron-transfer
barrier is significant for the couple. When the lower E° value
(+0.68 ± 0.02 V) is used for the ·/ ~ couple and the
self-exchange rates for the FeL33+/FeL32+ couples are taken
as -~109 M"1 s"1, the self-exchange rate for this ascorbate
couple (eq 7) is calculated from the Marcus cross-relation to
lie in the range 102— 104 M"1 s-1. This value for the exchange
rate is also consistent with that (104 M"1 s-1) calculated from
the Feni(cyt c)/HA" rate constant by using 1 X 103 M"1 s"1
for the cytochrome c self-exchange process.21 These consid-
erations thus implicate a much less than diffusion controlled
rate constant for the self-exchange of the HA"/HA· couple.
Undoubtedly part of the barrier to this electron-exchange
process derives from the differences in C-0 (and to a smaller
extent C-C) bond lengths in structures II and III in Figure
1. While eq 7 describes the self-exchange process relevant to
the bulk of reactions in which ascorbate is oxidized, the related
processes eq 8 and 9 may be the crucial ones at low and high
pH, respectively. As mentioned earlier, rate constants for
outer-sphere oxidation of H2A are available. Unfortunately
these cannot be interpreted in the Marcus framework without
an independent measure of E° for the 2 +·/ 2 couple. At
this time the E° evaluation cannot be made since the second
p of the ascorbate radical (i.e., for 2 +· ^ HA· + H+)
is not known. By contrast, for the A"·/A2" couple the E°
(+0.03 ± 0.02 V) is known but kinetic data are sparse. A
measure of the reactivity of this couple may, however, be
obtained through the cytochrome c results. With use of the
rate constant (/c3 = kl2 = 5 X 105 M~'s"1) for the A2"/Fera(cyt
c) reaction discussed above, k22 = 1 X 103 M"1 s"1 for Fen(cyt
c)/Fein(cyt c), and Ki2 = 7.9 X 103 (calculated from the E°
in Table II) a self-exchange rate ku ~ 105 M"1 s"1 is obtained
for eq 9. From these considerations the intrinsic electron-
transfer barrier for the A2"/A"· couple thus appears compa-
rable to or smaller than that for the HA"/A~· couple. This
conclusion is, of course, tentative since it rests on a single rate
constant. It is possible that some of the rate constants mea-
sured for reduction of the ascorbate radical (for example, that
for C1PMZ and ascorbate radical in ref 16) reflect a pathway
involving the A"·/A2" (as opposed to the HA·/HA") couple.
In general, however, these data cannot be interpreted since
no acid dependence (permitting disentangling of the HA· and
A"· reduction pathways) was carried out. Again, further
detailed kinetic studies could clarify some of these points.
In the previous paragraph, the reactivities of the ascorbate
radical/ascorbate couples were addressed. We turn now to
the process relevant to oxidation of the radical or reduction
of dehydroascorbic acid (eq 10). No studies of the kinetics
A"· + A ^ A + A'· (10)
of the reduction of dehydroascorbic acid appear to have been
made, and few bearing on the oxidation of the radical are
available. In fact the only quantitative information bearing
directly on this question are the rate constants for ascorbate
radical disproportionation at high pH (~105 M"1 s"1, de-
pending upon the medium11), an upper limit (5 X 102 M"1
s"1 n) for oxidation of ascorbate radical by 02, and that for
oxidation of A"· by ferricytochrome c. All of these seem rather
slow in light of the -0.14 ± 0.02 V E° for the A/A-· couple.
The cytochrome c results provide a useful basis for quantitative
comparisons. As mentioned earlier rate constants for oxidation
of A2- by Fein(cyt c) (k2 = 5 X 105 M"1 s"1) and for oxidation
of A · by Fein(cyt c) (k2 ~ 104 M"1 s"1) have been determined
(21) Gupta, R. K.; Koenig, S. H.; Redfield, A. G. J. Magn. Reson. 1972, 7,
66 and references cited therein.
4452 Inorg. Chem. 1981, 20, 4452-4453
(Scheme II). Oxidation of A2" involves the A"·/A2" couple
for which E° = +0.03 ± 0.02 V while oxidation of A"· involves
the A/A'· couple for which E° = -0.14 ± 0.02 V. Thus the
rate constant for the oxidation of the radical is smaller despite
the fact that the driving force for this reaction is greater than
for the oxidation of A2". As above, the self-exchange rate (eq
10) is estimated from the A'-/Fem(cyt c) reaction rate con-
stant; a rate constant k22 5 10'1 M'1 s'1 is thus obtained. Thus
we arrive at the conclusion that the A-·/A couple has a rather
high intrinsic electron-transfer barrier. While the evidence
is limited, sluggish redox properties for this couple might have
been predicted on the basis of the structures IV and VI shown
in Figure 1. The water-equilibrated form of dehydroascorbic
acid (whose properties Ball’s thermodynamic measurements
should reflect) has a bicyclic monoketo structure. It seems
likely that oxidation of A'·, a monocyclic species with extensive
radical delocalization over three >C=0 groups, yields initially
the monocyclic triketo form of dehydroascorbic acid V. Re-
arrangement of V to the stable form VI would then ensue.
Alternatively, rearrangement of the radical to a species more
strongly resembling VI could precede the electron-transfer step.
These alternative mechanisms are summarized in Schemes III
and IV where A(V) and A(VI) depict dehydroascorbic acid
in forms V and VI, respectively, and A-·' depicts a form of
ascorbate radical that is structurally similar to VI rather than
to V. Neither thermodynamic nor kinetic data bearing on the
equilibrium between structures V and VI are available; nor
is there any indication that A-·' exists. Thus the mechanism
of ascorbate radical oxidation cannot be clarified to any greater
extent at this time. It is worth noting, however, that, on the
basis of Scheme III or IV, neither the self-exchange estimate
obtained above nor the overall E° given in Table II is for the
simple process implied by eq 10 and that far from simple
kinetic behavior may be anticipated for reactions involving the
A'·/A couple.
Scheme 
A'· -* A(V) + e"
A(V) ^ A(VI)
Scheme IV
A · ^ A'·'
A"·' + e- — A(VI)
We mention one other complicating feature in passing.
While in most reactions involving ascorbate as a reductant eq
11 is likely to represent the initial step, it is worth reiterating
HA' + OX — HA- + RED (11)
that at pH 0-13 the stable form of the ascorbate radical is
the anionic species A"·. Thus above pH ~0 eq 11 is expected
to be followed by very rapid deprotonation of HA·, eq 4. With
HA· ^H+ + A"· (4)
fc-5
the assumption that fc_5 is nearly diffusion controlled (k-5 =
109 M"1 s'1), proton dissociation could proceed with k$ = 3.5
X 108 s'1; thus the lifetime of HA· may in some instances be
less than 10 ns. The rapid deprotonation of HA· introduces
an additional subtlety into the behavior of ascorbate systems:
while the rate of production of ascorbate radicals is related
to the properties (E°, kn) of the HA-/HA' couple, subsequent
reactions are determined by the properties of the A'·/A2' and
A/A'· couples discussed above. The rapid deprotonation of
HA· (as well as the sluggishness of the A'·/A couple) may
be of particular importance in the photochemical systems4
where it may provide a “switching” mechanism, effectively
slowing down some of the reactions that destroy reactive in-
termediates. In the thermal reactions reviewed in this paper
these interesting complications may also play an important,
if infrequently recognized, role. We hope that further detailed
studies of ascorbate reactions will lead to a fuller understanding
of these questions.
Acknowledgment. Helpful discussions with Drs. N. Sutin
and B. H. J. Bielski are gratefully acknowledged. This work
was performed under contract with the U.S. Department of
Energy and supported by its Office of Basic Energy Sciences.
Registry No. H2A, 50-81-7; A, 490-83-5.
Department of Chemistry Carol Creutz
Brookhaven National Laboratory
Upton, New York 11973Received December 31, 1980
Electron Correlation Effects in B4H10 Structures
Sir
The experimental geometry of B4H10 is well established1 as
a molecule of symmetry of styx topology 4012 in which
each of the two BH2 groups is joined by bridged hydrogens
to a singly bonded pair of BH units. An early proposal,2 also
of topology 4012, is the bis(diborane) structure, in which two
B2H5 units are joined by a single bond. Here, we address the
question, “What level of theoretical accuracy yields the ob-
served structure in preference to the bis(diborane) structure?”
In two previous calculations the observed structure is not
preferred. A 4-31G* study showed that bis(diborane) is more
stable than the observed structure by 16.5 kcal/mol. Neither
the addition of configuration interaction nor modification of
this basis set altered this conclusion.3 A study by Kleier,4 who
used the PRDDO method,5 yielded the gauche form of bis-
(diborane) as more stable than the trans form by about 1
kcal/mol, and more stable than the cis form by about 3
kcal/mol. The observed structure of (?> symmetry was about
5 kcal/mol less stable than the gauche form of bis(diborane).4
In the present study geometries were optimized6 with use
of a double-f basis set (3-21G),7 assuming symmetry for
the observed structure, C2h for íraní-bis(diborane), for
dí-bis(diborane), and C, for gaucAe-bis(diborane). The
gauche form optimized to C2 symmetry. These optimized
geometries (Table I) yielded the energies listed in Table II at
the levels 6-31G and 6-31G* (polarization added) and
Moller-Plesset (MP) corrections8 to third order9 for the 6-31G
basis (configuration interaction). In Table II we add the
(1) Lipscomb, W. N. “Boron Hydrides”; W. A. Benjamin: New York,
1963; and references therein.
(2) Pitzer, K. S. J. Am. Chem. Soc. 1945, 67, 1126.
(3) Yaniger, S. L; Shepard, R.; Simons, J.; Liebman, J. F., private com-
munication from J. Simons, Feb 1, 1980.
(4) Private communication.
(5) Halgren, T. A.; Kleier, D. A.; Hall, J. H., Jr.; Brown, L. D.; Lipscomb,
W. N. J. Am. Chem. Soc. 1978, 100, 6595.
(6) Binkley, J. S.; Whiteside, R. A.; Krishnan, R.; Seeger, R.; DeFrees, D.
J.; Schlegel, . B.; Topiol, S.; Kahn, L. R.; Pople, J. A. “Gaussian 80”;
Quantum Chemistry Program Exchange, Indiana University; Program
406.
(7) Binkley, J. S.; Pople, J. A.; Hehre, W. J. J. Am. Chem. Soc., 1980,102,
939.
(8) Metier, C.; Plesset, M. S. Phys. Rev. 1934, 46, 618.
(9) Pople, J. A.; Binkley, J. S.; Seeger, R. Int. J. Quantum Chem., Symp.
1976, 10, 1.
0020-1669/81/1320-4452S01.25/0 &copy; 1981 American Chemical Society